Ryan's question at Yahoo Answers (Eigenvalues of A^*A)

  • Context: MHB 
  • Thread starter Thread starter Fernando Revilla
  • Start date Start date
Click For Summary
SUMMARY

The discussion focuses on proving that all eigenvalues of the matrix product \( A^*A \) are nonnegative and that the matrix \( A^*A + I \) is invertible. It establishes that if \( \lambda \) is an eigenvalue of \( A^*A \), then \( \lambda \) can be expressed as \( \frac{||y||^2}{||x||^2} \), where \( y = Ax \) and \( x \neq 0 \), leading to the conclusion that \( \lambda \geq 0 \). Furthermore, it demonstrates that the determinant \( \det(A^*A + I) \neq 0 \), confirming the invertibility of \( A^*A + I \).

PREREQUISITES
  • Understanding of eigenvalues and eigenvectors in linear algebra
  • Familiarity with complex matrices and their properties
  • Knowledge of matrix determinants and invertibility criteria
  • Basic proficiency in linear transformations and their representations
NEXT STEPS
  • Study the properties of Hermitian matrices and their eigenvalues
  • Learn about the Singular Value Decomposition (SVD) of matrices
  • Explore the implications of the spectral theorem in linear algebra
  • Investigate the relationship between eigenvalues and matrix stability
USEFUL FOR

Mathematicians, students of linear algebra, and anyone involved in theoretical computer science or engineering applications requiring matrix analysis.

Fernando Revilla
Gold Member
MHB
Messages
631
Reaction score
0
Here is the question:

Let A be a mxn matrix. Prove that all eigenvalues of A*A are nonnegative and prove that A*A + I is invertible.

THANK YOU!

Here is a link to the question:

Linear Algebra Proof Help Please? - Yahoo! Answers

I have posted a link there to this topic so the OP can find my response.
 
Physics news on Phys.org
Hello Ryan,

Suppose $\lambda$ is an eigenvalue of $A^*A\in \mathbb{C}^{n\times n}$ then, there exists $x=(x_1,\ldots,x_n)^T\in\mathbb{C}^n$ such that $A^*Ax=\lambda x$. Multypling both sides by $x^*=(\overline{x_1},\ldots,\overline{x_n})$ we get $$x^*A^*Ax=(Ax)^*(Ax)=\lambda x^*x$$ But $y=Ax=(y_1,\ldots,y_m)^T$ and $x\neq 0$ (i.e. $||x||\neq0$), so $$\lambda=\frac{y^*y}{x^*x}=\frac{||y||^2}{||x||^2}\ge 0$$ On the other hand, $$\det (A^*A+I)=0\Leftrightarrow \det (A^*A-(-1)I)=0\Leftrightarrow -1\mbox{ is eigenvalue of }A^*A$$ This is a contradiction, so $\det(A^*A+I)\neq0$, as a consequence $A^*A+I$ is invertible.

If you have further questions, you can post them in the http://www.mathhelpboards.com/f14/ section.
 
Last edited:

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K